60 MIN · DEEP READ
Claude Code · Workflow · tmux
Sr Economic Analyst · Analytics EngineerTurning messy datainto decisionsworth shipping.
I'm Chris Pachulski: I turned a hobby into a career. Co-founded MTGBAN, a Magic: The Gathering arbitrage engine that grew to 500+ subscribers, while spending 7+ years shipping analytics pipelines across 5 industries for Ad.Net, Mozilla, SPINS, and Providencia. Now I'm a Senior Economic Analyst at Wizards of the Coast, working on the game I built a business around.
Python, SQL, R
Protection from Data
This creature's power is equal to the number of ETL pipelines you control, and its toughness is equal to that number plus 1.
"Short of leg. Long of opinion. The toddlers and dachshunds are brothers in arms and rule my life."
*/*+1
01 / about
Data done right feels like the lights
finally coming on.
Seven years keeping data layers honest. Pipelines that don't silently drop rows, warehouses that still compose six months in, models that survive past the second production deploy. SQL and Python daily; R when the problem actually needs a statistician.
Five industries before this: government contracting, adtech, browser telemetry, CPG retail panel, equity research. Now collectibles economics at Wizards. Stack rotates. Discipline doesn't.
Outside of it: my daughter, my son, three dachshunds, the perennial heartbreak of the Toronto Maple Leafs, and tinkering with Magic: the Gathering collections.
roleSr Economic Analyst · Analytics Engineer
currentWizards of the Coast, Magic: The Gathering
experience7+ yrs shipping data systems · 5 industries
previouslyThe Providencia Group · Ad.Net · Mozilla · Spins LLC · Consumer Edge Research
educationMBA, Hofstra · BA Classics + Biology, Creighton
locationSeattle, WA · Remote
02 / skills // SELECT * FROM tools ORDER BY depth DESC;
Here's where I can help.
I work in and mentor SQL and Python daily, and hold a strong nostalgia for R, the language I reach for when the problem needs a real statistician. Looker and BigQuery are where I spend most of my warehousing hours.
#
Tool
Where I use it
Years
01
SQL ●
analytics, warehousing, reporting
8y
02
Python ●
pipelines, reporting libraries, APIs
7y
03
R
forecasting, statistical analysis
7y
04
BigQuery
warehouse design, cost controls
5y
05
Looker
LookML modeling, exec dashboards
5y
06
Docker
reproducible reporting, CI
5y
featured
MTGBAN. I co-founded a $1.2M arbitrage engine for Magic: the Gathering.
500+ paying customers use our aggregated pricing to buy, sell, and hedge across a fragmented collectibles market. I designed the BigQuery warehouse from scratch, and the entire pipeline runs on Digital Ocean droplets + the R tidyverse.
Time-series ML forecasts price movements. A Go dashboard surfaces it all.
$1.2M ARR 500+ paying subs 24/7 since 2017
dag: mtgban_daily_pricing · 03:00 UTC
◉● run
scrape.tcgplayer
r · tidyverse · rvest
│
◎○ queued
scrape.cardkingdom
r · async http
│
◈○ queued
bigquery.stage
sql · partition by dt
│
◆○ queued
forecast.prophet
r · time series ml
│
✦○ queued
notify.subscribers
500+ paying users
03 / case studies // where I've delivered value
A curated list of the work I'd show you over coffee.
Projects where I've built something that outlasted me, or moved a number that mattered. Click any row for the challenge, the shape of the solution, and the numbers.
01.
Wizards of the Coast: MTG Economics & Analytics Platform
Econometrics and internal tooling for Magic: The Gathering, covering pricing, cannibalization, POS weighting, and revenue reconciliation across channels and set cycles.
02.
The Providencia Group: ORR National Call Center Platform
Data + reporting platform for a federal children's-safety hotline: trafficking, abuse, runaway, and safety case reporting for the Office of Refugee Resettlement.
03.
MTGBAN: Collectibles Pricing & Forecasting Platform
Eight years of continuous shipping. 38+ production R scripts (~26K LOC) running a multi-TCG pricing, forecasting, and distribution platform for 500+ paying subscribers.
04.
Mozilla Firefox: Ad Performance Analytics
Real-time ad performance dashboards + funnel analysis for a top-5 browser.
05.
Ad.Net: Centralized Python Reporting Library
One Python library replaced dozens of one-off scripts across the BI team.
06.
card-ops: Credit Card Portfolio AI Agent
Claude Code skill that evaluates card offers, compares cards, scans statements, and optimizes a personal credit-card portfolio, all with personal financial data kept local-only.
07.
session-sounds: Identity for Parallel AI Sessions
Every Claude Code / Codex session gets a named terminal tab + a unique notification sound, deduplicated across concurrent sessions, so you can run five in parallel and always know which one finished.
08.
clarity: Plan Maintenance for AI Agents
Scans an agent's plans/ directory, classifies each plan by staleness, extracts the decisions worth keeping into DECISIONS.md + memory, archives or prunes the rest.
09.
dream: Memory Consolidation for AI Agents
Audits dual-store memory (MCP semantic store + file-based memory/*.md), merges duplicates, decays stale entries, flags sticky memories nobody reinforces, and repairs the MEMORY.md index.
10.
ghscope: GitHub Repo Intelligence CLI
Point it at any GitHub repo, get a signal-dense scorecard: review coverage, reviewer concentration, bus factor, commit velocity, release cadence, first-timer retention, and the odds your PR will merge.
11.
Jira → Airflow DAG Generator
Auto-extract Jira issues, orchestrate tracking + reporting, surface timely insight for stakeholders.
12.
Python Utility Toolkit
A batteries-included library for my day-job: GCP, Salesforce, SharePoint, Genesys API, all in one import.
13.
Consumer Edge: Brand / NAICS Auto-Tagging Pipeline
Automated my own job. 23 R scripts, ~33K LOC, replacing manual brand classification with a scraping + rule-based NLP pipeline feeding a credit-card transaction panel used by banks.
04 / resume // git log --oneline --graph career
Seven years of shipping.
A git log of the work. Most recent at top, origin/HEAD marked. Financial services, digital marketing, CPG, tech, plus my own startup since 2017.
commit c0ffee1 · (HEAD)
Senior Economic Analyst
Wizards of the Coast
Nov 2025 to present
- Partner with product strategy, sales, finance, and economics to support Magic: The Gathering growth
- Build scalable reporting to manage revenue across channels, projects, and set cycles
- Deep-dive profitability analyses around pricing, inventory, and revenue recognition
- Reconcile revenue numbers alongside economists and financial analysts; explain findings with clear visuals
commit b92d81a
Senior Data Analyst
The Providencia Group
Sep 2024 to Nov 2025
- End-to-end reporting pipelines for the National Call Center across Genesys, Salesforce, SharePoint, SuccessKPI
- Rebuilt NCC reporting into a modular, version-controlled Python pipeline
- Dynamic staffing models using occupancy, shrinkage, and contact pacing
commit 7ec4a09
Senior Business Intelligence Analyst
Ad.Net
Dec 2021 to Sep 2024
- Looker dashboards for campaign + ad group performance
- Centralized Python library for reporting, querying, and exploration
- Jira API templates and scripts for automated report generation
- Salesforce + SimilarWeb integrated to uncover opportunities
- Docker, Anaconda, ClickHouse, PostgreSQL, R/H2O for forecasting
commit 3a1b7d2
Senior BI Reporting Specialist (Consultant)
Mozilla Firefox
Apr 2024 to Aug 2024
- Looker dashboards for real-time digital ad performance
- Funnel analysis optimizing conversion rates
- Automated reporting, reducing manual effort significantly
- Ad-hoc analyses on audience behavior and ad spend
commit f0c2ec1
Data Engineer
Spins LLC
Mar 2021 to Dec 2021
- Automated testing scripts ensuring data accuracy across platforms
- Optimized data operations, sourcing, and software performance
- Enhanced QA processes through automated test frameworks
commit 84e917b
Data Analyst
Consumer Edge Research
Sep 2019 to Mar 2021
- Credit card transaction data: automated brand entry from 15→300 / week
- Daily calls with financial institutions for onboarding and reviews
- Custom Looker dashboards for external client presentations
- Improved NLP brand tagging: model accuracy up 20%
commit a1f3e4c
Co-Founder
MTGBAN
2017 to 2025
- Hobby-turned-business: built a Magic: The Gathering arbitrage engine while working day-jobs
- 500+ paying customers using aggregated pricing
- Drove over $1.2M annual revenue for Magic: the Gathering products
- Designed and implemented BigQuery warehouse from the ground up
05 / writing // 50 posts
Field notes from the warehouse.
Practical write-ups: R, Python, SQL, and the occasional Docker-assisted life hack. I write them as I solve them.
Swift · tvOS · AVPlayer
Swift Was the Fourth Binding: Building the Emerald Exchange's Native Apple App
Rust · Plex · Transcoding
Watch Opens Plex — Until It Didn't: The Emerald Exchange Becomes the Player
Python · Genealogy · Playwright
"[not found]" Beats a Guess: A Six-Nation Family Tree Built From Primary Records Only
React · TypeScript · Family
One Link to Run the House: The Handbook I Built Before the Second Baby Came
Python · Zotero · Philosophy
The Canon Was the Scaffolding: 298 Philosophers, a Homelab, and One Question About Opaque AI
Claude Code · Docs · AI Tooling
The Docs Said Yes, the Code Said No: Building Evergreen as a Reflex, Not a Scanner
Rust · yt-dlp · Claude Code
"Infeasible" Was the AI Giving Up: Porting yt-dlp to Rust
Python · Mortgage Math · Claude Code
Math Correctness First: How Mortgage-Ops Forbids the LLM from Owning a Number
LATEST · MAY 2026
React · Plex · Claude Code
The Emerald Exchange: One Household, One Bookmark, and the Specialists I Got to Stop Being
Obsidian · Claude Code · YouTube
A Wiki That Earns Its Keep: Five Article Types, a Stop Hook, and a YouTube Intake Pipeline
Claude Code · Personal Tools
Adding a Brain to a Fork: career-ops, card-ops, and the Compiled-Context Pattern
HOME THEATER · S01E11
Plex · Usenet
At Home Media Server
4KHDRDOLBYUSENET
Claude Code · Codex · Windows
Four Terminals, Four Sounds: A Tab-Naming and Notification System for Parallel AI Sessions
28 MIN · DEEP READ
Claude Code · Econometrics · AI Tooling
Building an Autonomous Research Loop: The Stack, The Rationale, and What I Borrowed From Karpathy and Feynman
Claude Code · AI Tooling
Memory Hygiene for Long-Running AI Work: Anti-Stickiness, Dreams, and Plan Clarity
Python · Microsoft 365
Automating Microsoft 365 in Python Without an Azure App Registration
Python · Smartsheet
Streamlining Smartsheet with smartsheet_utils: A pandas-First Python Wrapper
Python · Windows
Production Python on Windows Task Scheduler: The Dual-Logging Pattern
Python · Salesforce · Genesys
The IDs Don't Match: Cross-System Reconciliation Between Genesys and Salesforce
Python · Genesys
Classifying Call-Center Agents with the Genesys API
R · Web Scraping
Pet Shop Monitoring with R
BigQuery · R
Efficient DB Structure and Cost Management in BigQuery
Python · SharePoint
Streamlining SharePoint with sharepoint_utility
SQL · BigQuery
Advanced SQL Techniques and Analytics
Python · Salesforce
Salesforce Data Management with Python and SOQL
R · API
Card Kingdom’s API Analysis
R · Python
R & Python — Comprehensive Set Up
Python · SQL
Automation for Social Advertising Data Management with Python and SQL
Python · RTB
Win Rate Analysis and Optimization for Blind RTB Bidding with Python and SQL
R · Python
R vs Python — Google Sheets
Bash · macOS
Bash Fun: Full On Sync & Setup For Mac
R · Analytics
Leveraging R for Advanced Client Revenue Analytics
Google Sheets · R
Inventory Acquisition with Google Sheets, Apps Script, SQL, and R
Bash
Bash Fun: Download Folder File Organizer
R · Market
In-Depth Product-Level Analysis Using R for Advanced Market Insights
R · Analytics
Advanced Analytics and Revenue Optimization with R
R · Advertising
Advanced Advertising Analytics with R: Unlocking Data-Driven Insights
R · SQL
Buying Support for Journey’s End Game Store Through R, SQL, and Communication
R · Python
R vs Python — Lazy vs Non-Lazy Evaluation
Python · SQL
Advanced Traffic Flow Analysis and Data Management with Python and SQL
R · Python
R vs Python — Package Management
R · GCP
MTGBAN — Newspaper Updater: R, Google Cloud Platform, BigQuery
Python · Gmail
Google Analytics and Gmail Automation with Python
ClickHouse · Py · R
ClickHouse: An In-Depth Overview and Integration with Python and R
R · Twitter
R & Twitter — Automation
R · Docker
How I Automated Booking a Baby Hospital Tour Using R and Docker
R · MTG
Arbitrage in Magic: The Gathering — a Primer
R · Blockchain
Racing Ahead: Data-Driven Insights into ZED RUN
R · MTG
R — Trading Card Market Analytics and Automated Reporting
06 / contact
Let's go.
Hiring for analytics, BI, or data-engineering work? Need help designing a warehouse from scratch, or rescuing one that's on fire? I answer every genuine email.
Python, SQL, R
Protection from Data
This creature's power is equal to the number of ETL pipelines you control, and its toughness is equal to that number plus 1.
"Short of leg. Long of opinion. The toddlers and dachshunds are brothers in arms and rule my life."
*/*+1